Context

How our agent, Möbius automated a Core ML port in ~12 h (vs. 2 weeks), hit 0.99998 parity, and made it 3.5× faster, while staying on the CPU.

NVIDIA product manager highlights the FluidAudio CoreML SDK enabling on‑device ASR on Apple devices; Parakeet‑TDT‑0.6B‑v3 runs at full speed for apps like Voice Ink, Spokenly, Slipbox, and Whisper Mate.

Profiles Moonshine AI’s offline transcription and translation models, featuring insights from Fluid Inference’s CEO on edge AI.

Our CEO joins Intel on AI to explore how Fluid Inference is moving large models from cloud data centers onto everyday devices, and why that shift matters for privacy, control, and cost.

A build journal for porting a PyTorch diarization stack to Core ML - covering ONNX graph surgery, ANE-friendly quantization, and a streaming scheduler that keeps labels under a second on Apple silicon.

Breaks down our joint work with Intel to land whisper-large-v3-turbo, qwen-3, and phi-4-mini on the NPU, including compiler patches, INT4 weight packing, and Windows packaging for AI PC developers.

Fluid Inference's work with Intel has unlocked a new tier of performance and privacy for customers looking to deploy transformer models on local devices.

Looks at why local-first products lag cloud AI: hardware fragmentation, tooling gaps, and distribution hurdles--and shares how we're closing each to make shipping offline AI boring.

Connects the dots from the early days of software portability to today's edge AI stack, and lays out the runtime, tooling, and business shifts needed for intelligence everywhere.